Various behaviors are often referred to as "hippocampal dependent" or "cerebellar dependent." In reality, brain function largely involves interactions between brain systems, although these interactions are difficult to study. The abundance of forebrain projections to cerebellum highlights the central prominence of interactions between forebrain and cerebellum. The practical and conceptual advantages of trace eyelid conditioning represent an opportunity to study forebrain-cerebellum interactions in the context of a well-defined learned behavior. Delay eyelid conditioning engages the cerebellum relatively directly and does not require forebrain structures. In contrast, trace eyelid conditioning is disrupted by lesions of the cerebellum, hippocampus and medial prefrontal cortex (mPFC). A prominent theory asserts that cerebellum cannot learn with trace inputs, that forebrain structures activate cerebellar inputs during the silent trace-interval, and that these inputs engage normal cerebellar learning mechanisms to acquire appropriate trace responses. The key test of this theory - to record from the mossy fiber inputs to cerebellum activated by hippocampus and mPFC - has not been possible because these mossy fibers had not been identified. The proposed studies extend preliminary findings that have 1) confirmed with mossy fiber stimulation that cerebellum cannot learn with trace inputs, and 2) identified the mossy fibers essential for trace conditioning. The experiments will be completed as a prelude to recording in vivo from the mossy fibers essential for trace conditioning and from the mossy fibers activated directly by the tone conditioned stimulus. The auditory responses of these cells and the learning-dependent activity that develops with trace conditioning will be characterized. Reversible inactivation of hippocampus or mPFC can then be used to identify essential sources of input that drive the learning-dependent responses. Finally, using stimulation of mossy fibers, the sufficiency of these learning dependent responses to support cerebellar learning of appropriate trace responses will be tested.